Premiere Pro text to speech: AI voice-over for your timeline
Premiere Pro turns speech into text for captions, but it does not turn text into speech. Adobe offers speech generation in its separate Firefly app; inside Premiere itself, a voice-over still has to come from somewhere else.
SuperClonica gives you natural narrators, character voices and 31 languages, free to try online, plus a free engine that voices a whole script into WAV files on your own computer. This guide shows the workflow from script to a mixed, captioned timeline in Premiere.
What Premiere can and cannot do
A generated voice is just another dialogue clip for Premiere, so every audio tool you already use works on it.
- Premiere Pro has Speech to Text for captions and transcripts, not text to speech.
- Adobe’s Firefly app can generate speech, as a separate service with its own account and plans.
- Premiere imports WAV and MP3, so a voice made anywhere drops onto the timeline.
- SuperClonica adds more than 170 catalog voices, 31 languages and a free engine that runs offline.
Voice a line free online
On the page of any voice, type a line of 30 to 100 characters, for example “Here is everything you need to know in under a minute.”, choose the language and press Voice it. In a few seconds you can listen and download the recording, with no account.
Bring the voice into Premiere
Generate keyframes again after you move a clip: each run replaces the old ones.
- Import the WAV file into your project and place it on its own audio track.
- In the Essential Sound panel, mark the voice as Dialogue and the music as Music.
- Turn on Ducking for the music and generate keyframes: the music dips under the voice automatically.
- Run Speech to Text on the voice track to get captions that match it word for word.
A whole script at once
Put the script into script.txt, one scene or paragraph per line, install the free engine with python -m pip install superclonica-cpu and voice every line into a numbered file: from superclonica import Superclonica; tts = Superclonica(); [tts.save(f'{i:03}.wav', tts.synthesize(line.strip(), 'F2', 'en', normalize=True)[0]) for i, line in enumerate(open('script.txt', encoding='utf-8'), 1) if line.strip()]. Import the folder, and the files land in timeline order at the same loudness, −18 LUFS. Change the script, voice that line again, and replace one clip instead of re-recording the whole take.
Voices for videos you publish
For a monetized channel, an ad or a client video, use the ten stock voices, M1 to M5 and F1 to F5, free under the OpenRAIL-M license, or your own voice trained in the SuperClonica studio. Character voices in the catalog are fan-made: fine for fan videos and parodies, not for commercial work.